[
https://issues.apache.org/jira/browse/HBASE-9668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13779572#comment-13779572
]
chunhui shen commented on HBASE-9668:
-------------------------------------
I don't think we should apply small scan for AggregationClient.
The function of small scan:
1.Reduce RPC, but it won't take affect in AggregationClient since we call it
through coprocessor
2.Using position read when reading data block from HDFS. Position read is
better if scan range is small, and seek+read is better if scan range is large.
Although the response is small, the scan range may be large(It means it will
scan a lot of data)
> Apply small scan to Scan's used by AggregationClient
> ----------------------------------------------------
>
> Key: HBASE-9668
> URL: https://issues.apache.org/jira/browse/HBASE-9668
> Project: HBase
> Issue Type: Improvement
> Reporter: Ted Yu
> Assignee: Ted Yu
> Attachments: 9668-v1.txt
>
>
> The Scan objects used to construct AggregateRequest's for aggregates
> supported by AggregationClient qualify as small scan because response from
> each region is small.
> We should utilize small scan for better performance.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
